Sei sulla pagina 1di 8


 8     Jeff  McLeod  

Is Ball Milling An Innovative Technique For the Production of Zn From ZnO?

Department of Physics
Goucher College
1021 Dulaney Valley Road
Baltimore, MD 21204, USA

Faculty Supervisor: Dr. Ali Bakhshai


The process of mechanical alloying using ball milling transfers mechanical energy to reactants in
powder form, causing the particle size of the reactant powders to be reduced until defects in the
lattice structure of the reactants are created. For reactions of sufficient exothermicity, this
facilitates a complete mechanochemical reaction through self-heat propagating synthesis (SHS).
The oxide reduction reaction of ZnO with Al, which yields pure Zn as a product, cannot be
induced using ball milling alone because of its low exothermicity. This study used a systematic
combination of ball milling and annealing through heat treatment in order to induce the reaction.
Parameters tested were milling time, annealing time, and annealing temperature with the purpose
of establishing the importance of each of these variables in inducing a complete reaction in the
sample. The completeness of the reaction was determined using x-ray diffraction analysis and
inspection with an optical microscope. Results confirmed that neither ball milling nor heat
treatment could induce the reaction individually; only ball milling followed by annealing could
cause the reaction to take place. An optimal milling time and annealing temperature were also
established. This study suggests that using ball milling in conjunction with heat treatment can
produce Zn from ZnO in a less costly, more efficient, easier, and less wasteful manner than
traditional methods.


Ball milling is a materials synthesis technique that is a key process in many applications under
the umbrella of mechanical alloying. In use since the 1970s during Chakurov’s study of
combination reactions[2], ball milling has become an innovative technique in the fields of
materials synthesis and mechanochemistry. It has specific applications in the production of
metallic coatings, nanocomposite materials, and alloys, often implicit in inducing oxide
reduction reactions[1].
Mechanical alloying involves an alternative for a chemical process using a transfer of
mechanical energy. Ball milling accomplishes this by catalyzing a chemical reaction in a
controlled environment through transfer of kinetic energy. During ball milling, reactant powders
are mixed inside a hardened steel vial containing several stainless steel balls. The vial is shaken
rapidly inside a ball mill, thus causing the balls inside to collide with each other as well as the
surface of the container. These rapid collisions decrease the size of the reactant particles until
defects in the crystal lattice structure of the particles themselves are created. These defect areas,
called active sites, facilitate the process of solid state diffusion by which the kinetic energy of the
steel balls is transferred to the powder, causing the desired alloying reaction to take place. By
this time, a thin layer of powder has formed on the balls and on the surface of the container,
Verge  8     McLeod  2  

bringing the active sites in close proximity to each other. Depending on the exothermicity of the
reaction, the reaction then propagates throughout the vial in a gradual process or through self-
heat propagating synthesis (SHS), a rapid process.
For certain reactions of low exothermicity, particularly copper oxide reduction reactions,
loose powder ignition causes the powder to react gradually throughout the milling time[2].
However, if a sufficient amount of heat is released, SHS can be induced. During SHS, the heat
released at each localized active site is transferred to a nearby site, thereby causing a rapid chain
reaction to proliferate through the vial. This chain reaction is called ignition, and typically
happens in less than three seconds. A variable of particular interest in the process of ball milling
is the time after the beginning of milling that ignition takes place; the ignition time tends to vary
depending on the constituent reactant particles, the mass of reactant powder, and the mechanical
energy supplied by the balls.
Although ball milling is particularly useful in reactions where SHS occurs within the vial,
research shows that it can also be used in conjunction with additional heat treatment for reactions
of low exothermicity or endothermic reactions, where SHS does not occur[1]. In the oxide
reduction reaction of ZnO with Al, ball milling is particularly useful because it causes lattice
defects to form, thereby facilitating a solid state diffusion process that can be induced using heat
treatment in a furnace at much lower temperatures than traditional chemical processes. Thermal
dissociation of ZnO occurs above 1900 °C[4], and chemical methods of production such as
pyrometallurgy processing and electrowinning require comparable temperatures as well as
expensive chemical solvents that can be harmful to the environment[3]. Since Zinc has important
industrial uses, including the production of brass for batteries, corrosion-resistant steel plating in
the aerospace industry, skin treatment products, and energy-saving windows, ball milling in
conjunction with heat treatment is an innovative process for the production of zinc because it is
less costly, less wasteful, and requires much lower temperatures than traditional methods[3].

Materials  and  Methods  

The experimental process was completed incrementally in five primary steps. First, the correct
amounts of each reactant powder had to be measured out and placed in the vials in a
LABCONCO controlled atmosphere glove box under minimal gauge pressure, in order to
minimize the possibility of oxygen contamination that could disrupt the reaction. The reactant
powders were then milled in a SPEX 8000 mixer mill for a set time period. Upon completion of
the milling period, the composite powder was pressed into a one-half inch disk using a hydraulic
press in order to facilitate the annealing process. Finally, heat treatment was performed in a tube
furnace and samples were analyzed using one or more of the following instruments: an optical
microscope, Goucher College’s Atomic Force Microscope, and the Towson University Physics
Department’s x-ray diffractometer.
The reactants used were zinc oxide powder and aluminum powder supplied by Alfa
Aesar. Three grams of the reactant mixture were used in each trial, with the individual amounts
of each compound determined by stoichiometry of the following reaction:

3ZnO + 2Al  Al2O3 + 3Zn, ΔH = -632 kJ/mol

The mixture was milled by placing it in a SPEX cylindrical hardened steel vial with five stainless
steel balls of different sizes. For four trials, three large (one-half inch, 8.25 gram) and two
medium (three-eighths inch, 3.48 gram) balls were used, manufactured by Small Parts Inc. For
Verge  8     McLeod  3  

the remaining sixteen trials, two large balls and three medium balls were used. This change was
made in order to increase the kinetic energy transferred to the powder, since smaller balls travel
faster within the vial due to the increased mean free path between collisions which results in
fewer, but more energetic, collisions taking place. The variable of particular interest in the
milling process was the time of milling; this took on a value of ten minutes to seventy-five
minutes over all the trials. Milling time was determined to be an important factor in the degree of
success of the combined ball milling and annealing process, since the formation of active sites
within the powder depends on the number and severity of collisions that take place within the
The ball milling technique described above is accompanied by several disadvantages in
addition to the aspects that make it useful for experimental materials synthesis research. Due to
the high number, intensity and frequency of collisions within the vial, small impurities will
inevitably be introduced into the sample that are trace elements of the hardened steel vial and
balls, and these impurities can cause experimental data to be skewed. However, one of the main
reasons why the ball milling step is so useful in the experimental process is that it is dependable
and produces consistent results; we expect that milling a set amount of a powder mixture for a
set duration with the same intensity (ball mass) will cause a similar process to take place within
the vial each time a trial is performed. For the purpose of this research, it was worthwhile to
analyze the dependability of the ball milling technique more thoroughly, and this analysis was
performed prior to the ZnO reduction reaction experiment by performing several trials using a
reaction of high exothermicity:

Bi2O3 + 2Al  Al2O3 +2Bi, ∆H = -1102 kJ/mol

The bismuth oxide reduction reaction was exothermic enough to cause an SHS process to take
place within the vial. Six trials were performed with the same powder mass and the same milling
intensity of three large balls, with the objective of reproducing a consistent time of ignition
across each trial. An Omega Engineering Inc. thermocouple was attached to the vial during
milling in order to record the temperature of the vial and occurrence of ignition was indicated by
a sharp increase in temperature as depicted in Figure 1. The results of this preliminary
investigation, depicted in Table 1, indicated the reproducibility of the ball milling technique; the
small standard deviation in ignition time across the six trials was clearly indicative of a
consistent and predictable process taking place within the vial and we can reasonably conclude
from this data that the technique should be as dependable for a set amount of a different
Continuing with the ZnO investigation, after the milling cycle was completed, the powder
was removed from the vial and pressed using a Carver Laboratory hydraulic press at a pressure
of eleven metric tons. Samples were then placed in heat-resistant ceramic boats (Coors Ceramic
Company) and placed in a Lindberg digital tube furnace for annealing in an argon gas flow. Up
to three samples at a time were exposed to each heat treatment cycle.
Analysis of the milled and annealed samples focused on several properties which were
indicative of the completeness of the reaction. Surface morphology of samples was analyzed
using Goucher College’s QScope Atomic Force Microscope; however, after performing many
trials and noting the properties shared by the most successful samples the completeness of the
reaction could be judged using a National DC2-155 Digital optical microscope or naked eye
observation. The most accurate method of determining the chemical composition of each sample
Verge  8     McLeod  4  

involved analysis using the Broker AXS diffractometer in collaboration with Dr. Raj Kolagani at
Towson University. This way, the presence of each expected compound (Zn, ZnO, Al, Al2O3)
was detected and used to indicate whether the sample had completely or partially reacted
(Figures 3 & 4).
The experimental process that we performed was a systematic approach involving
observation of the effects of changing a particular variable while holding others constant over
several trials. This approach allowed us to determine which variables were most important in
causing a complete reaction to take place within the sample. The three main variables studied
were milling time, annealing time, and annealing temperature. Successive trials were performed
in order to minimize each of these variables while still producing a reacted sample.

Results and Discussion

This investigation was carried out using a systematic approach designed to resolve the role of
key experimental parameters involved in catalyzing the desired reaction in the sample.
Parameters tested were milling time, annealing time, and annealing temperature. The expectation
was that milling of the sample for a sufficient period of time would allow for the minimization of
the other variables of interest – that is, significant reductions in the temperature and length of the
heat treatment. These reductions are desirable because they reflect the overall usefulness of the
ball milling technique in the production of Zn from ZnO: waste reduction, increased efficiency
and reduced cost. The results are depicted in Table 2, with the first five trials delineated from the
rest because of the difference in milling intensity: these trials used three large balls and two
medium balls for milling, while the rest of the trials were performed using two large balls and
three medium balls in order to increase the mechanical energy transferred in each collision.
The samples which exhibited signs that a complete reaction had taken place were those
milled for at least forty-five minutes and annealed for four hours at 425°C. As shown in Table 2,
these optimal parameters were determined by holding one variable constant while testing several
values of a different variable. More focus was placed on determining the interdependent
relationship between milling time and annealing temperature than on determining a lower bound
for the annealing time, although the samples which were annealed for less than four hours were
at most only partly reacted.
The first set of trials processed samples with a milling time of 10, 20, and 30 minutes at
an annealing temperature of 400°C. Analysis of these samples determined that a complete
reaction had not taken place. In addition to these three trials, control trials were performed in
order to confirm the relationship between the milling process and the annealing process; that is,
that the latter complements the former in inducing a complete reaction in the sample. One control
was a sample milled for sixty minutes with no subsequent heat treatment, and the other was a
sample that was annealed for four hours at 400°C but did not undergo ball milling. Neither of
these samples showed even slight signs that a reaction had taken place. This is because the
reaction of ZnO with Al is not exothermic enough to trigger SHS during the milling process, so
while some isolated reactions may take place the reaction does not propagate throughout the vial.
Additional heat treatment is needed to trigger the SHS process. The reader may question, at this
point, why ball milling followed by annealing causes the sample to react but annealing without
milling is not sufficient, or why the ball milling and annealing processes could not be performed
in reverse order. The answer has to do with the effects of the ball milling process itself. As
described in earlier sections, the ball milling process causes lattice defects in the crystal structure
Verge  8     McLeod  5  

of the ZnO molecules. These areas, where the chemical bonds are broken, become “active sites”
that facilitate the process of solid state diffusion between molecules. The solid state diffusion
process causes the reaction to take place, and the additional heat treatment ensures that the SHS
process will occur, by which the heat released in each localized reaction induces a reaction in
neighboring regions. Without the active sites formed during the milling process, thermal
dissociation of ZnO requires a much higher temperature. The formation of active sites within the
crystal structure of the ZnO molecules allows the SHS process to occur at much lower
temperatures – in this experiment, as low as 425°C.
Subsequent trials were performed at 425°C and a new range of milling times. The most
successfully reacted samples analyzed came from this group, in which an optimal value for
milling time was also established: samples milled for at least 45 minutes or up to 60 minutes had
the most complete reactions. Lower milling times produced partly reacted samples while milling
samples for longer than 60 minutes did not cause a noticeable increase in the completeness of the
After annealing, the samples were analyzed to determine the completeness of the
reaction. This analysis was performed using x-ray diffraction, which indicated the presence of
both the reactants and products within the sample (Figures 3 & 4). After performing several trials
and noting which conditions allowed for the most complete reactions to take place, however,
preliminary indicators of reaction completeness could be determined using naked-eye
observation of the sample or an optical microscope. In samples where a reaction had taken place,
the zinc was easily noticeable in the form of silver wiry strands or droplets on the surface of the
sample. Samples that did not undergo a complete reaction remained dark greyish black with little
or no indication that zinc was present.
As a result of this investigation, an optimal milling time and annealing temperature were
established – 60 minutes of milling at 425°C produced samples with the most complete
reactions. The explanation for these results lies in the combined microscopic effects of the ball
milling and annealing processes. The idea is that in order to extract Zn from ZnO, chemical
bonds must be broken requiring high temperatures. Ball milling causes small defects in the
lattice structure of ZnO and Al molecules, allowing for the oxide reduction reaction to take place
in small, localized regions during the milling process. In this way, trace amounts of Zn are
present at these active sites after ball milling is completed and before annealing is performed.
During heat treatment, once the melting point of Zn – 419.58°C – is reached, this small amount
of Zn melts and soon afterwards reaches phase equilibrium with ZnO at 419.6°C[3]. At this point,
an unknown reaction involving energy exchange occurs between the ZnO and melted Zn that
causes SHS to occur during annealing. The equilibrium temperature of 419.6°C agrees closely
with the temperature at which we observed the most successful reactions, 425°C. Future research
can be done which examines the relevant phase shifts in greater detail to determine how exactly
the SHS process is initiated by the energy exchange between ZnO and liquid Zn.


This investigation has demonstrated the consistency and dependability of the ball milling
technique as a method of materials synthesis as well as its usefulness, when combined with
moderate heat treatment, in producing pure Zn from ZnO via an oxide reduction reaction. Using
x-ray diffraction analysis as well as naked-eye and optical microscope observations, analysis
showed a dependent relationship between milling time, annealing temperature, and annealing
Verge  8     McLeod  6  

time in causing the sample to react fully during annealing. This method is an inexpensive, less
time consuming and less wasteful alternative to Zn mining or chemical methods of Zn
production. Further investigations can more closely hone in on optimum milling and annealing
time, in order to make the process even more efficient.


[1] Samuels, Margaret. “The Production of Si from SiO2 via the Ball-Milling Technique.”
Proceedings of the National Conference On Undergraduate Research (2008).

[2] Takacs, L. “Self-Sustaining Reactions Induced by Ball Milling: An Overview.” International

Journal of Self-Propagating High-Temperature Synthesis 18.4 (2009): 276-82.

[3] Welcome to Zinc World - Your Web Resource for All Things Zinc. International Zinc
Association. Web. 21 Jan. 2011. <>.

[4] Wriedt, H. A. “The O-Zn (Oxygen-Zinc) System.” Bulletin of Alloy Phase Diagrams 8.2
(1987): 166-76. Print.
Verge  8     McLeod  7  

Sample  #   Duration   Ignition  

(s)   time  (s)  
102   684   336  

103   582   333  

104   594   367  

105   615   305  

106   615   355  

107   615   354  

Figure 1: Temperature vs. time for one trial in the
Mean   342  ±  22    
Bi2O3 investigation. Ignition occurs at 367 seconds.
ignition   sec  

Table 1: Results for the Bi2O3


Sample Milling Annealing Annealing

# time time (hrs) Temp.
(min) (°C)
8 0 4 400
9 10 4 400
10 20 4 400
11 30 4 400
12 60 0 ~
16 30 4 425
17 20 4 425
18 45 4 425
19 60 4 425
20 75 4 425
21 10 4 425
22 55 4 425
23 65 4 425
24 75 4 425
Figure 2: Completeness of reactions: a cross indicates 25 30 4 400
no reaction, diamond partly reacted, circle full reaction 26 45 4 400
27 60 4 400
28 45 1 400
29 45 1 375
30 45 2 375
31 60 4 375

Table 2: Results for the ZnO investigation.

The shaded rows represent a different
milling intensity.
Verge  8     McLeod  8  

Figure 3: XRD data for a fully reacted sample. A Figure 4: XRD data for a partly reacted
high-intensity Zn peak is present. sample. The peaks for Zn and ZnO are
almost the same.